Evaluating the risk of phospholipidosis using a new multidisciplinary pipeline approach

•We generated a toxicophore model for phospholipidosis (PLD) inducers.•We selected compounds by virtual screening to be tested in biological assays.•Biological assays based on different mechanisms can lead to different PLD effects.•For one compound all the in silico approaches failed in predicting the PLD effect.

Phospholipidosis (PLD) is an undesirable potential side-effect of drugs, and cationic amphiphilic drugs (CADs) represent the main class of PLD inducers. A CADs toxicophore has been recently proposed, although the CADs definition is far from being trivial. In this work we derive a three-dimensional CADs toxicophore (here named PLD-phore) using a molecular interaction field approach, and test its suitability to discriminate between PLD inducers and non-inducers in a virtual screening approach. Ten commercially available compounds predicted to be PLD inducers and non-inducers based on their similarity to the PLD-phore were experimentally tested for PLD induction using two cell-based in vitro assays (fluorescent lipid uptake, activity of secreted lysosomal β-hexosaminidase). When a positive effect was observed, the PLD induction was also confirmed by transmission electron microscopy. Two exceptions to the general statement about CADs and PLD induction were detected and discussed, and for one compound the cell-based in-vitro assays lead to different outcomes.

Graphical abstractFigure optionsDownload full-size imageDownload as PowerPoint slide